use atomic decrement rather than cas in pthread_exit thread count
[musl] / crt / x86_64 /
drwxr-xr-x   ..
-rw-r--r-- 633 Scrt1.s
-rw-r--r-- 584 crt1.s
-rw-r--r-- 73 crti.s
-rw-r--r-- 41 crtn.s